arabianights said:
some say edward witten, but much of his theories is unproven and much less useful, imho
And not even capitalizing proper nouns qualifies for what again ?
Fundamental science is not like sport or horse races. Even failed approaches can have propaedeutic value. Although Newton did not mean it in as a compliment, researchers do "sit on the shoulders of giants". Therefore, there is not much value in trying to classify speculative ideas in fundamental physics. There is much more value in trying to understand for oneself the established concepts, sitting on the latest giants and making one's own mind as to which speculation will be the most fruitful in the future.
This being said, and because Ed Witten is being bullied here, I will take a couple of examples related but not restricted to him. There are a great many people to praise for the revolutionary concepts they put forward. Their proposal may not be as straightforward to grasp as fancy colorful science-fiction movie graphical animations, but they also go much deeper and wider.
So take Witten's most cited paper over the last decade. Dec 2003 "Perturbative gauge theory in twistor space". As it turns out, Roger Penrose has been insisting that we need to use his twistor formulation. His arguments are not just technical, but also extend to a full philosophical discussion and picture. Maybe because Penrose's background is in general relativity, the particle physics community has not been listening much. But after Witten's paper in 2003, the twistor methods have boomed all the way into Monte-Carlo generators essentials for the analysis of data coming out of the LHC. Penrose's proposal has only very few rivals in elegance, breadth and depth. I would urge anybody interested to know more to read his "Road to Reality", which is a true masterpiece in the communication of modern science.
One possible candidate for a revolutionary shift in our picture of the fundamental laws is Alain Connes' non-commutative geometry. In the beginning of the previous century, Einstein proposed to use non-euclidean geometries. We are talking about the same kind of shift : opening the possibilities for geometries which are qualitatively different. Connes likes to tell the story of a well known theoretician who walked out of a seminar Connes gave long ago. When they recently and randomly met aboard a train, the well known theoretician approached Connes and starts asking him questions. Surprised, Connes reminds him of the time he walked out of the seminar. The answer : "but I recently saw Witten hold your book in a library".
This last example should also illustrate the influence Ed Witten has on modern physics. Whether one agrees with him or not, the influence cannot be denied.
Alain Connes has also written the most beautiful speculative paper I have ever read, with Carlo Rovelli :
Von Neumann Algebra Automorphisms and Time-Thermodynamics Relation in General Covariant Quantum Theories
